Xiao Wei, an audio algorithm engineer at Chinese tech giant Tencent, has dedicated roughly five years to refining sound through artificial intelligence (AI), specializing in noise elimination and audio purification through AI noise reduction technology. This innovation offers users an enhanced listening experience, ensuring clarity and authenticity.

As an algorithm engineer within the emerging field of AI, his work commences with the meticulous collection of noise samples, a fundamental yet labor-intensive step crucial for machine learning. Equipped with headphones and a recorder, he has traveled far and wide to amass a broad array of noise data.

"We have collected around 300 types of noise," he told Xinhua News Agency. The team's acoustic laboratory, designed to simulate real-world noise environments, is a square room equipped with eight speakers, creating an extreme testing environment essential for recording high-quality training data. Over the past five years, Xiao Wei has lost count of how many lines of code he has written and how many times he has tested and iterated algorithms.

His profession belongs to the realm of AI engineering and technology. Since 2019, the Ministry of Human Resources and Social Security has been steadily introducing new digital professions in AI. So far, China officially recognizes 106 digital careers.

The new sphere

Zhai Yujia is an AI digital human trainer at China's leading video-sharing platform Kuaishou. Her primary responsibility is training a digital persona called Guan Xiaofang. "We train not only her IQ but also her EQ, enabling her to better perceive and understand humans, consequently facilitating more interesting conversations and interactions," she told China Central Television (CCTV).

The role of an AI digital human trainer is among the new digital professions recently listed by the Ministry of Human Resources and Social Security, categorized under the profession of AI trainers. Guan Xiaofang is a product driven by large-scale AI models, which are typically built with billions, or even trillions, of parameters.

In the context of AI applications, virtual digital humans have become an important direction in recent years, leading to the rise of different digital personas—from anchors and narrators to intelligent customer service agents and digital avatars. Now, with the rapid integration of large-scale AI models, digital human products are leaping toward higher levels of intelligence.

"At the beginning of the training, we need to prepare hundreds of thousands of data points, including both manually written data and basic data generated by models," Zhai said. "We continuously optimize and supplement the training data to create a positive cycle, making Guan Xiaofang smarter and wiser."

Liu Hongyu, AI Product Manager at China-based tech company Huawei, is at the forefront of a new profession: generative AI system operation. This role focuses on applying large-scale AI models across different industries. For example, Liu and his team are developing a customs knowledge assistant application, demonstrating the practical implementation of this innovative technology.

To create a customs knowledge assistant, they collected tens of thousands of pieces of information regarding customs and international trade from around the world, annotated them into a high-quality dataset, and trained a large language model specialized in customs knowledge Q&A using a general model.

Then, by fine-tuning the prompting methods and optimizing the model workflow, they continuously improved the accuracy of its responses, ultimately building a convenient and useful customs knowledge assistant application.

Over the past year, large-scale AI models have been continuously upgraded, and generative AI has set off a new tech wave. Industries urgently need generative AI applications for empowerment, leading to the emergence of the new profession of generative AI system application operators.

"From the perspective of professional evaluation standards, generative AI, despite its short history, has rapidly spawned professions," Fan Wei, a researcher with the Chinese Academy of Personnel Science, told CCTV. "The government promptly announced related professions to guide and regulate the orderly and rational development of this technology and its practitioners. According to our estimates, a new profession can generally generate jobs for 300,000 to 500,000 people."

From campus to company

Figures from the China Academy of Information and Communications Technology show that as of late 2022, China's digital economy had reached a scale of 50.2 trillion yuan ($7 trillion), accounting for 41.5 percent of GDP. The swift progress has led to an increasingly significant shortage of AI talent. Current estimates indicate that the talent gap now exceeds 5 million people.

Earlier this year, the Ministry of Human Resources and Social Security, along with nine other departments, issued an action plan for accelerating the cultivation of digital talent to support the development of the digital economy (2024-26), aiming to align with the needs of digital industrialization and the digitization of industries, intending to increase the effective supply of digital talent within approximately three years.

Many enterprises and universities have already moved into action. Since 2021, leading Internet company Meituan has collaborated with Tsinghua University, launching dozens of research cooperation projects in intelligent unmanned systems, language intelligence and the digital economy. They have also introduced long-term courses on the applications of big data technology.

Zhang Hongchao, who recently completed his Ph.D. in microelectronics and solid-state electronics at Beihang University in Beijing, secured an early job offer from a microelectronics company in Beijing, where he had previously interned for over two years as a device research and development engineer while simultaneously working on his academic projects.

Interning at the memory chip developer gave him both considerable experience and a deep understanding of corporate needs, and allowed him to master related technologies.

"My upcoming job aligns with my Ph.D. research topics," he told Xinhua. "All this allows me to quickly adapt to the work."

In the opinion of Li Xueming, Director of the Talent Strategy and Policy Research Office at the Chinese Academy of Personnel Science, digital talent plays a major role in technological innovation and is a key support for transforming economic development modes and promoting development.

"Accelerating the cultivation of digital talent, especially focusing on developing high-level digital engineers and highly skilled professionals, is both an immediate priority and a long-term strategy," he told Xinhua.

(Print edition title: Hi-Tech Horizons)

Copyedited by Elsbeth van Paridon

Comments to yuanyuan@cicgamericas.com